Think AI "knows" what it’s doing? Scientists say think again

2026-04-19 · Source: Artificial Intelligence News -- ScienceDaily · Field: Science & Research — Social Sciences & Behavioral Studies, Research Methodology & Innovation · Depth: Novice, medium

Summary

A new study by researchers from Iowa State University and Brigham Young University, published in *Technical Communication Quarterly*, investigates how news writers use human-like language, or anthropomorphism, when describing artificial intelligence. The study, titled "Anthropomorphizing Artificial Intelligence: A Corpus Study of Mental Verbs Used with AI and ChatGPT," analyzed the News on the Web (NOW) corpus, a dataset of over 20 billion words from English news articles. Contrary to common assumptions, the research found that news writers rarely use strongly anthropomorphic language, such as mental verbs like "think," "know," or "understand," when referring to AI or ChatGPT. When such terms do appear, their meaning often depends heavily on context, frequently describing basic requirements (e.g., "AI needs data") rather than implying human-like cognition or intent. The study highlights that anthropomorphism exists on a spectrum, with some uses subtly suggesting deeper capabilities while others are purely functional.

Key takeaway

For technical communicators and journalists reporting on AI, you should carefully consider your word choices to avoid unintentionally misleading readers about AI's capabilities. Recognize that terms like "needs" can be functional rather than anthropomorphic, but phrases implying "understanding" or "deciding" can blur the line between human and machine. Your language shapes public perception and accountability, so prioritize clarity and accuracy over casual anthropomorphism to set realistic expectations for AI systems.

Key insights

Anthropomorphic language in AI news coverage is less frequent and more nuanced than commonly assumed.

Principles

• Context dictates whether mental verbs applied to AI are anthropomorphic.
• Anthropomorphism in AI description exists on a spectrum.

Method

Researchers analyzed the News on the Web (NOW) corpus, focusing on the frequency of mental verbs paired with "AI" and "ChatGPT" in news articles.

In practice

• Analyze language context, not just word count, for anthropomorphism.
• Consider editorial guidelines like AP style for AI reporting.

Topics

• Anthropomorphism
• AI Language
• Mental Verbs
• News on the Web Corpus
• Public Perception of AI

Best for: Tech Journalist, AI Ethicist, Research Scientist

Related on AIssential

• Think AI ‘knows’ what it’s doing? Well, think again - Digital Journal — Journalists use anthropomorphic language for AI less frequently and more nuancedly than often assumed.
• we JUST figured out how AI thinks — Natural Language Autoencoders translate AI internal activations into human-readable text, enhancing interpretability and safety.
• When Machines 'Think': Anthropomorphization in the Discourse on AI - Conceptual Risks and Proposal for a Non-Anthropomorphic Lexicon for NLP in Portuguese — Anthropomorphic language in AI discourse creates conceptual misunderstandings and misrepresents system capabilities.
• Can AI Truly Understand Human Language? — AI models recognize language patterns and approximate meaning, but lack true human-like understanding.
• ChinAI #362: Chinese Encounters with "Artificial Challenged Intelligence" [人工智障] — AI's probabilistic nature often leads it to fabricate answers rather than admit ignorance, creating "Artificial Challenged Intelligence" and potential societal risks.
• What is Anthropic? — AI labs hold divergent philosophies on model agency, impacting design and perceived moral authority.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Intelligence News -- ScienceDaily.